Python 数据操作

import pandas as pd

data_name = '.data/1.csv'
df_csv = pd.read_csv(data_name) # excel-excel table-txt csv-csv
print(df_csv.head())

# 读取多个文件合并成一个文件
import glob
path = '.data/Bearing1_1'
# 合并多个数组
all_files = glob.glob(path+'/*.csv')
all_data = []
for filename in all_files:
    df = pd.read_csv(filename, index_col=None, header=0)
    all_data.append(df)
data_123_outerrace = pd.concat(all_data, axis=0, ignore_index=True)
print(data_123_outerrace)

data = df_csv
print(data.shape)           # 维度
print(data.dtypes)          # 数据类型
print(data.info)            # 基本信息
print(data.describe())      # 平均值等等
print(data.columns)         # 列名
print(data['Horizontal_vibration_signals'].unique())    # 某列一共有多少参数
print(data['Horizontal_vibration_signals'].values)      # 某列的所有值
print(data.head())            # 查看前五行
print(data.tail(3))            # 查看后五行/n=3:查看后3行

# 查找空值
print(data.isnull().any(axis=0))    # axis=0:列 / axis=1:行
print(data.isnull().any(axis=1))
# print(data.loc[data.isnull().any(axis=0)])      # 定位到缺失的
print(data.isnull().sum(axis=0))                # 求各变量缺失的数量

data1 = data.dropna()                       # 直接删除空值所在行
print(data1)
data2 = data.fillna(0, inplace=False)       # 定义空值为零
print(data2)
'''data2 = data.fillna(value={'Horizontal_vibration_signals':data['Horizontal_vibration_signals'].mode()[0],   # 用众数代替空值
                           'Vertical_vibration_signals':data['Horizontal_vibration_signals'].mean()},       # 用平均数代替空值
                    inplace=False)'''

print(df_csv.duplicated().any())               # 判断数据中是否有重复值
# data3 = data.drop_duplicates()              # 删除重复值
# 指定变量删除， First 保留第一个， inplace=True 对原数据进行修改
# data4 = data.drop_duplicates(subject=['Horizontal_vibration_signals'], keep='First', inplace=False)

data['index'] = data.index                              # 增加索引
print(data.head())
data['New'] = data['Horizontal_vibration_signals']      # 增加一列
print(data.head())
data3 = data.drop(columns='New')                        # 删除某列，注意要重新附值
print(data3.head())
data4 = data.drop(index=2)                              # 根据索引，删除特定行
print(data4.head())
data5 = data.drop(index=range(0, 16, 2))                # index=[1, 3 ,6]>>>删除特定几行
print(data5.head(20))

df_csv1 = pd.read_csv(data_name) # excel-excel table-txt csv-csv
# 过滤某个区间
data6 = df_csv1[(df_csv1['Horizontal_vibration_signals'] <= 1.0) & (df_csv1['Horizontal_vibration_signals'] >= -1.0)]
print(data6.head(30))
# 值排序:从小到大，一列值
print(data6.head(30).sort_values(by='Vertical_vibration_signals'))
# 从大到小
print(data6.head(30).sort_values(by='Vertical_vibration_signals', ascending=False))

# 数据汇总 data7 = df_csv1.groupby(by='Horizontal_vibration_signals')['Vertical_vibration_signals'].std()
print(df_csv1.describe())
print(df_csv1[['Horizontal_vibration_signals', 'Vertical_vibration_signals']].std())

# 数据标准化
import numpy as np
# 0-1标准化
def MaxMinNormalization(x):
    x = (x-np.min(x))/(np.max(x)-np.min(x))
    return x
x1 = MaxMinNormalization(df_csv1['Horizontal_vibration_signals']).reset_index()
print(x1)
# Z-score标准化
def ZescoreNormalization(x):
    x = (x-np.mean(x))/np.std(x)
    return x
x2 = ZescoreNormalization(df_csv1['Horizontal_vibration_signals'])
print(x2)
# 数据储存
# x2.to_excel('./Test_data/x2.xls', sheet_name='Horizontal_vibration_signals')